This invention relates generally to a system for transmitting data along a drillstring, and more particularly to a system for transmitting data through a drillstring by modulation of intermediate-frequency acoustic carrier waves.
Deep wells of the type commonly used for petroleum or geothermal exploration are typically less than 30 cm (12 inches) in diameter and on the order of 2 km (1.5 miles) long. These wells are drilled using drillstrings assembled from relatively light sections (either 30 or 45 feet long) of drill pipe that are connected end-to-end by tool joints, additional sections being added to the uphole end as the hope deepens. The downhole end of the drillstring typically includes a drill collar, a weight assembled from sections of relatively heavy lengths of uniform diameter collar pipe having an overall length on the order of 300 meters (1000 feet). A drill bit is attached to the downhole end of the drill collar, the weight of the collar causing the bit to bite into the earth as the drillstring is rotated from the surface. Sometimes, downhole mud motors or turbines are used to turn the bit. Drilling mud or air is pumped from the surface to the drill bit through an axial hole in the drillstring. This fluid removes the cuttings from the hole, provides a hydrostatic head which controls the formation gases, provides a deposit on the wall to seal the formation, and sometimes provides cooling for the bit.
Communication between downhole sensors of parameters such as pressure or temperature and the surface has long been desirable. Various methods that have been tried for this communication include electromagnetic radiation through the ground formation, electrical transmission through an insulated conductor, pressure pulse propagation through the drilling mud, and acoustic wave propagation through the metal drillstring. Each of these methods has disadvantages associated with signal attenuation, ambient noise, high temperatures and compatibility with standard drilling procedures.
The most commercially successful of these methods has been the transmission of information by pressure pulse in the drilling mud. However, attenuation mechanisms in the mud limit the transmission rate to less than 1 bit per second.
This invention is directed towards the acoustical transmission of data through the metal drillstring. The history of such efforts is recorded in columns 2-4 of U.S. Pat. No. 4,293,936, issued Oct. 6, 1981, of Cox and Chaney. As reported therein, the first efforts were in the late 1940's by Sun Oil Company, which organization concluded there was too much attenuation in the drillstring for the technology at that time. Another company came to the same conclusion during this period.
U.S. Pat. No. 3,252,225, issued May 24, 1966, of E. Hixon concluded that the length of the drill pipes and joints had an effect on the transmission of energy up the drillstring. Hixon determined that the wavelength of the transmitted data should be greater than twice and preferably four times the length of a section of pipe.
In 1968 Sun Oil tried again, using repeaters spaced along the drillstring and transmitting the best frequency range, one with attenuation of only 10 dB/1000 feet. A paper by Thomas Barnes et al., "Passbands for Acoustic Transmission in an Idealized Drillstring", Journal of Acoustical Society of America, Vol. 51, No. 5, 1972, pages 1606-1608, was consulted for an explanation of the field-test results, which were not totally consistent with the theory. Eventually, Sun went back to random searching for the best frequencies for transmission, an unsuccessful procedure.
The aforementioned Cox and Chaney patent concluded from their interpretation of the measured data obtained from a field test in a petroleum well that the Barnes model must be in error, because the center of the passbands measured by Cox and Chaney did not agree with the predicted passbands of Barnes et al. The patent uses acoustic repeaters along the drillstring to ensure transmission of a particular frequency for a particular length of drillpipe to the surface.
U.S. Pat. No. 4,314,365, issued Feb. 2, 1982, of C. Petersen et al discloses a system similar to Hixon for transmitting acoustic frequencies between 290 Hz and 400 Hz down a drillstring.
U.S. Pat. No. 4,390,975, issued Jun. 28, 1983, of E. Shawhan, noted that ringing in the drillstring could cause a binary "zero" to be mistaken as a "one". This patent transmitted data, and then a delay to allow the transients to ring down before transmitting subsequent data.
U.S. Pat. No. 4,562,559, issued Dec. 31, 1985, of H. E. Sharp et al, uncovered the existence of "fine structure" within the passbands; e.g., "such fine structure is in the nature of a comb with transmission voids or gaps occurring between teeth representing transmission bands, both within the overall passbands." Sharp attributed this structure to "differences in pipe length, conditions of tool joints, and the like." The patent proposed a complicated phase shifted wave with a broader frequency spectrum to bridge these gaps.
The present invention is based upon a more thorough consideration of the underlying theory of acoustical transmission through a drillstring. For the first time, the work of Barnes et al, has been analyzed as a banded structure of the type discussed by L. Brillouin, Wave Propagation in Periodic Structures, McGraw-Hill Book Co., New York, 1946. The theoretical results of this invention have also been correlated to extensive laboratory experiments on scale models of the drillstring, and the original data tape obtained from Cox and Chaney's field test has been reanalyzed. This analysis shows that Cox and Chaney's measurements contain data which, in fact, is in excellent agreement with the theoretical predictions of Barnes and this invention; that Sharp misinterpreted the cause of the fine structure; and that the ringing and frequency limitations cited by Shawhan and Hixon are easily overcome by signal processing.
FIG. 1 shows some of the results of the new analysis of the data recorded by Cox and Chaney. This FIGURE is a plot of the power amplitude versus frequency of the transmitted signal. The theoretical boundaries between the passbands and the stopbands are shown by the vertical dotted lines. If this FIGURE is compared to FIG. 1 in Cox and Chaney's patent significant and obvious differences can be noted. These are attributable to error in Cox and Chaney's signal analysis. Furthermore, FIG. 1 of this invention also shows the "fine structure" of Sharp et al. From the analysis of this invention we now know that this fine structure is caused by echoes bouncing between opposite ends of the drillstring, the number of peaks being correlated to the number of sections of drillpipe. A theoretical calculation of this field test was used to produce FIG. 2. All of the phenomena important to the transmission of data in the drillstring is represented in this calculation. These theoretical results accurately predict the location of the passbands and the fine structure produced by the echo phenomena.